Model tuning or prompt Tuning? a study of large language models for clinical concept and relation extraction

OBJECTIVE

To develop soft prompt-based learning architecture for large language models (LLMs), examine prompt-tuning using frozen/unfrozen LLMs, and assess their abilities in transfer learning and few-shot learning.

METHODS

We developed a soft prompt-based learning architecture and compared 4 strategies including (1) fine-tuning without prompts; (2) hard-prompting with unfrozen LLMs; (3) soft-prompting with unfrozen LLMs; and (4) soft-prompting with frozen LLMs. We evaluated GatorTron, a clinical LLM with up to 8.9 billion parameters, and compared GatorTron with 4 existing transformer models for clinical concept and relation extraction on 2 benchmark datasets for adverse drug events and social determinants of health (SDoH). We evaluated the few-shot learning ability and generalizability for cross-institution applications.

RESULTS AND CONCLUSION

When LLMs are unfrozen, GatorTron-3.9B with soft prompting achieves the best strict F1-scores of 0.9118 and 0.8604 for concept extraction, outperforming the traditional fine-tuning and hard prompt-based models by 0.6 ∼ 3.1 % and 1.2 ∼ 2.9 %, respectively; GatorTron-345 M with soft prompting achieves the best F1-scores of 0.8332 and 0.7488 for end-to-end relation extraction, outperforming other two models by 0.2 ∼ 2 % and 0.6 ∼ 11.7 %, respectively. When LLMs are frozen, small LLMs have a big gap to be competitive with unfrozen models; scaling LLMs up to billions of parameters makes frozen LLMs competitive with unfrozen models. Soft prompting with a frozen GatorTron-8.9B model achieved the best performance for cross-institution evaluation. We demonstrate that (1) machines can learn soft prompts better than hard prompts composed by human, (2) frozen LLMs have good few-shot learning ability and generalizability for cross-institution applications, (3) frozen LLMs reduce computing cost to 2.5 ∼ 6 % of previous methods using unfrozen LLMs, and (4) frozen LLMs require large models (e.g., over several billions of parameters) for good performance.

Identifying social determinants of health from clinical narratives: A study of performance, documentation ratio, and potential bias

OBJECTIVE

To develop a natural language processing (NLP) package to extract social determinants of health (SDoH) from clinical narratives, examine the bias among race and gender groups, test the generalizability of extracting SDoH for different disease groups, and examine population-level extraction ratio.

METHODS

We developed SDoH corpora using clinical notes identified at the University of Florida (UF) Health. We systematically compared 7 transformer-based large language models (LLMs) and developed an open-source package - SODA (i.e., SOcial DeterminAnts) to facilitate SDoH extraction from clinical narratives. We examined the performance and potential bias of SODA for different race and gender groups, tested the generalizability of SODA using two disease domains including cancer and opioid use, and explored strategies for improvement. We applied SODA to extract 19 categories of SDoH from the breast (n = 7,971), lung (n = 11,804), and colorectal cancer (n = 6,240) cohorts to assess patient-level extraction ratio and examine the differences among race and gender groups.

RESULTS

We developed an SDoH corpus using 629 clinical notes of cancer patients with annotations of 13,193 SDoH concepts/attributes from 19 categories of SDoH, and another cross-disease validation corpus using 200 notes from opioid use patients with 4,342 SDoH concepts/attributes. We compared 7 transformer models and the GatorTron model achieved the best mean average strict/lenient F1 scores of 0.9122 and 0.9367 for SDoH concept extraction and 0.9584 and 0.9593 for linking attributes to SDoH concepts. There is a small performance gap (∼4%) between Males and Females, but a large performance gap (>16 %) among race groups. The performance dropped when we applied the cancer SDoH model to the opioid cohort; fine-tuning using a smaller opioid SDoH corpus improved the performance. The extraction ratio varied in the three cancer cohorts, in which 10 SDoH could be extracted from over 70 % of cancer patients, but 9 SDoH could be extracted from less than 70 % of cancer patients. Individuals from the White and Black groups have a higher extraction ratio than other minority race groups.

CONCLUSIONS

Our SODA package achieved good performance in extracting 19 categories of SDoH from clinical narratives. The SODA package with pre-trained transformer models is available at https://github.com/uf-hobi-informatics-lab/SODA_Docker.

Selective-Area-Grown PbTe-Pb Planar Josephson Junctions for Quantum Devices

Planar Josephson junctions are predicted to host Majorana zero modes. The material platforms in previous studies are two-dimensional electron gases (InAs, InSb, InAsSb, and HgTe) coupled to a superconductor such as Al or Nb. Here, we introduce a new material platform for planar JJs, the PbTe-Pb hybrid. The semiconductor, PbTe, was grown as a thin film via selective area epitaxy. The Josephson junction was defined by a shadow wall during the deposition of superconductor Pb. Scanning transmission electron microscopy reveals a sharp semiconductor-superconductor interface. Gate-tunable supercurrents and multiple Andreev reflections are observed. A perpendicular magnetic field causes interference patterns of the switching current, exhibiting Fraunhofer-like and SQUID-like behaviors. We further demonstrate a prototype device for Majorana detection wherein phase bias and tunneling spectroscopy are applicable.

Generative large language models are all-purpose text analytics engines: text-to-text learning is all your need

OBJECTIVE

To solve major clinical natural language processing (NLP) tasks using a unified text-to-text learning architecture based on a generative large language model (LLM) via prompt tuning.

METHODS

We formulated 7 key clinical NLP tasks as text-to-text learning and solved them using one unified generative clinical LLM, GatorTronGPT, developed using GPT-3 architecture and trained with up to 20 billion parameters. We adopted soft prompts (ie, trainable vectors) with frozen LLM, where the LLM parameters were not updated (ie, frozen) and only the vectors of soft prompts were updated, known as prompt tuning. We added additional soft prompts as a prefix to the input layer, which were optimized during the prompt tuning. We evaluated the proposed method using 7 clinical NLP tasks and compared them with previous task-specific solutions based on Transformer models.

RESULTS AND CONCLUSION

The proposed approach achieved state-of-the-art performance for 5 out of 7 major clinical NLP tasks using one unified generative LLM. Our approach outperformed previous task-specific transformer models by ∼3% for concept extraction and 7% for relation extraction applied to social determinants of health, 3.4% for clinical concept normalization, 3.4%-10% for clinical abbreviation disambiguation, and 5.5%-9% for natural language inference. Our approach also outperformed a previously developed prompt-based machine reading comprehension (MRC) model, GatorTron-MRC, for clinical concept and relation extraction. The proposed approach can deliver the "one model for all" promise from training to deployment using a unified generative LLM.

LncRNA MACC1-AS1 induces gemcitabine resistance in pancreatic cancer cells through suppressing ferroptosis

Pancreatic ductal adenocarcinoma (PDA) mortality is primarily attributed to metastasis and chemotherapy resistance. In this research, the long non-coding RNA MACC1-AS1 was studied, playing a significant role in regulating lipid oxidation processes. This regulation could further lead to the inhibition of ferroptosis induced by chemotherapeutic drugs, making it a contributing factor to gemcitabine resistance in PDA. In both gemcitabine-resistant PDA patients and mouse models, the elevated expression level of MACC1-AS1 in the tumors was noted. Additionally, overexpression of MACC1-AS1 in pancreatic cancer cells was found to enhance tolerance to gemcitabine and suppress ferroptosis. Proteomic analysis of drug-resistant pancreatic cells revealed that overexpressed MACC1-AS1 inhibited the ubiquitination degradation of residues in the protein kinase STK33 by MDM4. Furthermore, its accumulation in the cytoplasm activated STK33, further activating the ferroptosis-suppressing proteins GPX4, thereby counteracting gemcitabine-induced cellular oxidative damage. These findings suggested that the long non-coding RNA MACC1-AS1 could play a significant role in the ability of pancreatic cancer cells to evade iron-mediated ferroptosis induced by gemcitabine. This discovery holds promise for developing clinical therapeutic strategies to combat chemotherapy resistance in pancreatic cancer.

[Effects of cerium oxide nanoenzyme-gelatin methacrylate anhydride hydrogel in the repair of infected full-thickness skin defect wounds in mice]

Objective: To investigate the effects of cerium oxide nanoenzyme-gelatin methacrylate anhydride (GelMA) hydrogel (hereinafter referred to as composite hydrogel) in the repair of infected full-thickness skin defect wounds in mice. Methods: This study was an experimental study. Cerium oxide nanoenzyme with a particle size of (116±9) nm was prepared by hydrothermal method, and GelMA hydrogel with porous network structure and good gelling performance was also prepared. The 25 μg/mL cerium oxide nanoenzyme which could significantly promote the proliferation of human skin fibroblasts and had high superoxide dismutase activity was screened out. It was added to GelMA hydrogel to prepare composite hydrogel. The percentage of cerium oxide nanoenzyme released from the composite hydrogel was calculated after immersing it in phosphate buffer solution (PBS) for 3 and 7 d. The red blood cell suspension of mice was divided into PBS group, Triton X-100 group, cerium oxide nanoenzyme group, GelMA hydrogel group, and composite hydrogel group, which were treated with corresponding solution. The hemolysis of red blood cells was detected by microplate reader after 1 h of treatment. The bacterial concentrations of methicillin-resistant Staphylococcus aureus (MRSA) and Escherichia coli were determined after being cultured with PBS, cerium oxide nanoenzyme, GelMA hydrogel, and composite hydrogel for 2 h. The sample size in all above experiments was 3. Twenty-four 8-week-old male BALB/c mice were taken, and a full-thickness skin defect wound was prepared in the symmetrical position on the back and infected with MRSA. The mice were divided into control group without any drug intervention, and cerium oxide nanoenzyme group, GelMA hydrogel group, and composite hydrogel group applied with corresponding solution, with 6 mice in each group. The wound healing was observed on 3, 7, and 14 d after injury, and the remaining wound areas on 3 and 7 d after injury were measured (the sample size was 5). The concentration of MRSA in the wound exudation of mice on 3 d after injury was measured (the sample size was 3), and the blood flow perfusion in the wound of mice on 5 d after injury was observed using a laser speckle flow imaging system (the sample size was 6). On 14 d after injury, the wound tissue of mice was collected for hematoxylin-eosin staining to observe the newly formed epithelium and for Masson staining to observe the collagen situation (the sample size was both 3). Results: After immersion for 3 and 7 d, the release percentages of cerium oxide nanoenzyme in the composite hydrogel were about 39% and 75%, respectively. After 1 h of treatment, compared with that in Triton X-100 group, the hemolysis of red blood cells in PBS group, GelMA hydrogel group, cerium oxide nanoenzyme group, and composite hydrogel group was significantly decreased (P<0.05). Compared with that cultured with PBS, the concentrations of MRSA and Escherichia coli cultured with cerium oxide nanoenzyme, GelMA hydrogel, and composite hydrogel for 2 h were significantly decreased (P<0.05). The wounds of mice in the four groups were gradually healed from 3 to 14 d after injury, and the wounds of mice in composite hydrogel group were all healed on 14 d after injury. On 3 and 7 d after injury, the remaining wound areas of mice in composite hydrogel group were (29±3) and (13±5) mm2, respectively, which were significantly smaller than (56±12) and (46±10) mm2 in control group and (51±7) and (38±8) mm2 in cerium oxide nanoenzyme group (with P values all <0.05), but was similar to (41±5) and (24±9) mm2 in GelMA hydrogel group (with P values both >0.05). On 3 d after injury, the concentration of MRSA on the wound of mice in composite hydrogel group was significantly lower than that in control group, cerium oxide nanoenzyme group, and GelMA hydrogel group, respectively (with P values all <0.05). On 5 d after injury, the volume of blood perfusion in the wound of mice in composite hydrogel group was significantly higher than that in control group, cerium oxide nanoenzyme group, and GelMA hydrogel group, respectively (P<0.05). On 14 d after injury, the wound of mice in composite hydrogel group basically completed epithelization, and the epithelization was significantly better than that in the other three groups. Compared with that in the other three groups, the content of collagen in the wound of mice in composite hydrogel group was significantly increased, and the arrangement was also more orderly. Conclusions: The composite hydrogel has good biocompatibility and antibacterial effect in vivo and in vitro. It can continuously sustained release cerium oxide nanoenzyme, improve wound blood perfusion in the early stage, and promote wound re-epithelialization and collagen synthesis, therefore promoting the healing of infected full-thickness skin defect wounds in mice.

[Research progress of long-chain non-coding RNA in lipid metabolism reprogramming in primary hepatocellular carcinoma]

Hepatocellular carcinoma (HCC) is the most common liver malignant tumor with complex pathogenesis and a poor prognosis. Metabolic reprogramming has been recognized as one of the important cancer markers, and the liver, as an important organ for lipid metabolism in the human body, plays an important role in the process of the occurrence and development of HCC. More and more evidence shows that long-chain non-coding RNA (lncRNA) can influence the lipid metabolism process by regulating key enzymes and transcription factors, as well as being involved in the occurrence and development of HCC. Therefore, explicating the mechanism of lncRNA in lipid metabolism reprogramming is conducive to providing new targets and strategies for the diagnosis and treatment and improving the prognosis of HCC patients. This article summarizes the latest research progress on the involvement of lncRNA in the reprogramming process of HCC lipid metabolism.

Heat stress-associated changes in the intestinal barrier, inflammatory signals, and microbiome communities in dairy calves

Recent studies indicate that heat stress pathophysiology is associated with intestinal barrier dysfunction, local and systemic inflammation, and gut dysbiosis. However, inconclusive results and a poor description of tissue-specific changes must be addressed to identify potential intervention targets against heat stress illness in growing calves. Therefore, the objective of this study was to evaluate components of the intestinal barrier, pro- and anti-inflammatory signals, and microbiota community composition in Holstein bull calves exposed to heat stress. Animals (mean age = 12 wk old; mean body weight = 122 kg) penned individually in temperature-controlled rooms were assigned to (1) thermoneutral conditions (constant room temperature at 19.5°C) and restricted offer of feed (TNR, n = 8), or (2) heat stress conditions (cycles of room temperatures ranging from 20 to 37.8°C) along with ad libitum offer of feed (HS, n = 8) for 7 d. Upon treatment completion, sections of the jejunum, ileum, and colon were collected and snap-frozen immediately to evaluate gene and protein expression, cytokine concentrations, and myeloperoxidase activity. Digesta aliquots of the ileum, colon, and rectum were collected to assess bacterial communities. Plasma was harvested on d 2, 5, and 7 to determine cytokine concentrations. Overall, results showed a section-specific effect of HS on intestinal integrity. Jejunal mRNA expression of TJP1 was decreased by 70.9% in HS relative to TNR calves. In agreement, jejunal expression of heat shock transcription factor-1 protein, a known tight junction protein expression regulator, decreased by 48% in HS calves. Jejunal analyses showed that HS decreased concentrations of IL-1α by 36.6% and tended to decrease the concentration of IL-17A. Conversely, HS elicited a 3.5-fold increase in jejunal concentration of anti-inflammatory IL-36 receptor antagonist. Plasma analysis of pro-inflammatory cytokines showed that IL-6 decreased by 51% in HS relative to TNR calves. Heat stress alteration of the large intestine bacterial communities was characterized by increased genus Butyrivibrio_3, a known butyrate-producing organism, and changes in bacteria metabolism of energy and AA. A strong positive correlation between the rectal temperature and pro-inflammatory Eggerthii spp. was detected in HS calves. In conclusion, this work indicates that HS impairs the intestinal barrier function of jejunum. The pro- and anti-inflammatory signal changes may be part of a broader response to restore intestinal homeostasis in jejunum. The changes in large intestine bacterial communities favoring butyrate-producing organisms (e.g., Butyrivibrio spp.) may be part of a successful response to maintain the integrity of the colonic mucosa of HS calves. The alteration of intestinal homeostasis should be the target for heat stress therapies to restore biological functions, and, thus highlights the relevance of this work.

[Research progress of non-coding RNA-encoding polypeptides in primary hepatocellular carcinoma]

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors, with rapid progression and a poor prognosis. More and more studies have shown that there are small open reading frames (sORFs) on the molecular sequences of a large number of non-coding RNAs (ncRNAs), which can encode conserved peptides that play an important role in controlling the occurrence and development of HCC. This article introduces the discovery, prediction, and validation methods of ncRNA-encoding polypeptides and reviews its research progress, with the aim of providing new targets and ideas for early-stage diagnosis, targeted therapy, and prognosis assessment of HCC.

Extracting Thyroid Nodules Characteristics from Ultrasound Reports Using Transformer-based Natural Language Processing Methods

The ultrasound characteristics of thyroid nodules guide the evaluation of thyroid cancer in patients with thyroid nodules. However, the characteristics of thyroid nodules are often documented in clinical narratives such as ultrasound reports. Previous studies have examined natural language processing (NLP) methods in extracting a limited number of characteristics (<9) using rule-based NLP systems. In this study, a multidisciplinary team of NLP experts and thyroid specialists, identified thyroid nodule characteristics that are important for clinical care, composed annotation guidelines, developed a corpus, and compared 5 state-of-the-art transformer-based NLP methods, including BERT, RoBERTa, LongFormer, DeBERTa, and GatorTron, for extraction of thyroid nodule characteristics from ultrasound reports. Our GatorTron model, a transformer-based large language model trained using over 90 billion words of text, achieved the best strict and lenient F1-score of 0.8851 and 0.9495 for the extraction of a total number of 16 thyroid nodule characteristics, and 0.9321 for linking characteristics to nodules, outperforming other clinical transformer models. To the best of our knowledge, this is the first study to systematically categorize and apply transformer-based NLP models to extract a large number of clinical relevant thyroid nodule characteristics from ultrasound reports. This study lays ground for assessing the documentation quality of thyroid ultrasound reports and examining outcomes of patients with thyroid nodules using electronic health records.

A Cholecystokinin Analogue Ameliorates Cognitive Deficits and Regulates Mitochondrial Dynamics via the AMPK/Drp1 Pathway in APP/PS1 Mice

BACKGROUND

There are no drugs on the market that can reverse or slow Alzheimer's disease (AD) progression. A protease-resistant Cholecystokinin (CCK) analogue used in this study is based on the basic structure of CCK, which further increases the stability of the peptide fragment and prolongs its half-life in vivo. We observed a neuroprotective effect of CCK-8L in APPswe/PS1dE9 (APP/PS1) AD mice. However, its corresponding mechanisms still need to be elucidated.

OBJECTIVE

This study examined CCK-8L's neuroprotective effects in enhancing cognitive impairment by regulating mitochondrial dynamics through AMPK/Drp1 pathway in the APP/PS1 AD mice.

METHODS

Behavioural tests are applied to assess competence in cognitive functions. Transmission electron microscopy (TEM) was performed to observe the ultrastructure of mitochondria of hippocampal neurons, Immunofluorescent staining was employed to assay for Aβ1-42, APP, Adenosine 5'-monophosphate (AMP)-activated protein kinase (AMPK) and dynamin-related protein1 (Drp1). CRISPR/Cas9 was utilized for targeted knockout of the CCKB receptor (CCKBR) in the mouse APP/PS1 hippocampal CA1 region. A model of lentiviral vector-mediated overexpression of APP in N2a cells was constructed.

RESULTS

In vivo, experiments revealed that CCK analogue and liraglutide significantly alleviated cognitive deficits in APP/PS1 mice, reduced Aβ1-42 expression, and ameliorated l damage, which is associated with CCKBR activation in the hippocampal CA1 region of mice. In vitro tests showed that CCK inhibited mitochondrial fission and promoted fusion through AMPK/Drp1 pathway.

CONCLUSIONS

CCK analogue ameliorates cognitive deficits and regulates mitochondrial dynamics by activating the CCKB receptor and the AMPK/Drp1 pathway in AD mice.

Mild photothermal therapy assist in promoting bone repair: Related mechanism and materials

Achieving precision treatment in bone tissue engineering (BTE) remains a challenge. Photothermal therapy (PTT), as a form of precision therapy, has been extensively investigated for its safety and efficacy. It has demonstrated significant potential in the treatment of orthopedic diseases such as bone tumors, postoperative infections and osteoarthritis. However, the high temperatures associated with PTT can lead to certain limitations and drawbacks. In recent years, researchers have explored the use of biomaterials for mild photothermal therapy (MPT), which offers a promising approach for addressing these limitations. This review provides a comprehensive overview of the mechanisms underlying MPT and presents a compilation of photothermal agents and their utilization strategies for bone tissue repair. Additionally, the paper discusses the future prospects of MPT-assisted bone tissue regeneration, aiming to provide insights and recommendations for optimizing material design in this field.

Ballistic PbTe Nanowire Devices

Disorder is the primary obstacle in the current Majorana nanowire experiments. Reducing disorder or achieving ballistic transport is thus of paramount importance. In clean and ballistic nanowire devices, quantized conductance is expected, with plateau quality serving as a benchmark for disorder assessment. Here, we introduce ballistic PbTe nanowire devices grown by using the selective-area-growth (SAG) technique. Quantized conductance plateaus in units of 2e2/h are observed at zero magnetic field. This observation represents an advancement in diminishing disorder within SAG nanowires as most of the previously studied SAG nanowires (InSb or InAs) have not exhibited zero-field ballistic transport. Notably, the plateau values indicate that the ubiquitous valley degeneracy in PbTe is lifted in nanowire devices. This degeneracy lifting addresses an additional concern in the pursuit of Majorana realization. Moreover, these ballistic PbTe nanowires may enable the search for clean signatures of the spin-orbit helical gap in future devices.

Tissue-resident alveolar macrophages reduce O3-induced inflammation via MerTK mediated efferocytosis

Lung inflammation, caused by acute exposure to ozone (O3) - one of the six criteria air pollutants - is a significant source of morbidity in susceptible individuals. Alveolar macrophages (AMØs) are the most abundant immune cells in the normal lung and their number increases following O3 exposure. However, the role of AMØs in promoting or limiting O3-induced lung inflammation has not been clearly defined. Here, we used a mouse model of acute O3 exposure, lineage tracing, genetic knockouts, and data from O3-exposed human volunteers to define the role and ontogeny of AMØs during acute O3 exposure. Lineage tracing experiments showed that 12, 24, and 72 h after exposure to O3 (2 ppm) for 3h all AMØs were tissue-resident origin. Similarly, in humans exposed to FA and O3 (200 ppb) for 135 minutes, we did not observe ~21h post-exposure an increase in monocyte-derived AMØs by flow cytometry. Highlighting a role for tissue-resident AMØs, we demonstrate that depletion of tissue-resident AMØs with clodronate-loaded liposomes led to persistence of neutrophils in the alveolar space after O3 exposure, suggesting that impaired neutrophil clearance (i.e., efferocytosis) leads to prolonged lung inflammation. Moreover, depletion of tissue-resident AMØ demonstrated reduced clearance of intratracheally instilled apoptotic Jurkat cells, consistent with reduced efferocytosis. Genetic ablation of MerTK - a key receptor involved in efferocytosis - also resulted in impaired clearance of apoptotic neutrophils followed O3 exposure. Overall, these findings underscore the pivotal role of tissue-resident AMØs in resolving O3-induced inflammation via MerTK-mediated efferocytosis.

TransFuser: Imitation With Transformer-Based Sensor Fusion for Autonomous Driving

How should we integrate representations from complementary sensors for autonomous driving? Geometry-based fusion has shown promise for perception (e.g., object detection, motion forecasting). However, in the context of end-to-end driving, we find that imitation learning based on existing sensor fusion methods underperforms in complex driving scenarios with a high density of dynamic agents. Therefore, we propose TransFuser, a mechanism to integrate image and LiDAR representations using self-attention. Our approach uses transformer modules at multiple resolutions to fuse perspective view and bird's eye view feature maps. We experimentally validate its efficacy on a challenging new benchmark with long routes and dense traffic, as well as the official leaderboard of the CARLA urban driving simulator. At the time of submission, TransFuser outperforms all prior work on the CARLA leaderboard in terms of driving score by a large margin. Compared to geometry-based fusion, TransFuser reduces the average collisions per kilometer by 48%.

CREB3L2 Regulates Hemidesmosome Formation during Epithelial Sealing

The long-term success rate of dental implants can be improved by establishing a favorable biological sealing with a high-quality epithelial attachment. The application of mesenchymal stem cells (MSCs) holds promise for facilitating the soft tissue integration around implants, but the molecular mechanism is still unclear and the general application of MSC sheet for soft tissue integration is also relatively unexplored. We found that gingival tissue-derived MSC (GMSC) sheet treatment significantly promoted the expression of hemidesmosome (HD)-related genes and proteins in gingival epithelial cells (GECs). The formation of HDs played a key role in strengthening peri-implant epithelium (PIE) sealing. Further, high-throughput transcriptome sequencing showed that GMSC sheet significantly upregulated the PI3K/AKT pathway, confirming that cell adhesion and HD expression in GECs were regulated by GMSC sheet. We observed that the expression of transcription factor CREB3L2 in GECs was downregulated. After treatment with PI3K pathway inhibitor LY294002, CREB3L2 messenger RNA and protein expression levels were upregulated. Further experiments showed that overexpression or knockdown of CREB3L2 could significantly inhibit or promote HD-related genes and proteins, respectively. We confirmed that CREB3L2 was a transcription factor downstream of the PI3K/AKT pathway and participated in the formation of HDs regulated by GMSC sheet. Finally, through the establishment of early implant placement model in rats, we clarified the molecular function of CREB3L2 in PIE sealing as a mechanical transmission molecule in GECs. The application of GMSC sheet-implant complex could enhance the formation of HDs at the implant-PIE interface and decrease the penetration distance of horseradish peroxidase between the implant and PIE. Meanwhile, GMSC sheet reduced the length of CREB3L2 protein expression on PIE. These findings elucidate the potential function and molecular mechanism of MSC sheet regulating the epithelial sealing around implants, providing new insights and ideas for the application of stem cell therapy in regenerative medicine.

Principal Component Analysis to Design Planning Target Volume in Oropharyngeal Cancers

PURPOSE/OBJECTIVE(S)

Standard translational shifts of the Clinical Target volume (CTV) to generate the Planning Target Volume (PTV) do not account for rotations. Head and neck positional misalignments derive in large part from rotations due to cervical spine arching and twisting in Cone Beam Computed Tomography (CBCT). Translational expansions do not track rotations, yielding coverage envelopes that unnecessarily overlap with adjacent structures. This work examines whether principal component analysis of the motion along all 6 degrees of freedom may be used to produce a more favorable PTV.

MATERIALS/METHODS

Seventy-five CBCTs of ten oropharyngeal cases were included. The records of couch shifts needed to align individual bony structures (C1-5, mandible and mastoid) between the planning image and CBCTs were recorded. A Principal Component Analysis of the shifts was used to generate an ellipsoid inflation of each CTV vertex along 6 degrees of freedom. The result was compared to a 3D ellipsoid based translational expansion, and to a described ellipsoid based vertex expansion along 6 degrees of freedom, with axes oriented in parallel to the treatment reference frame.

RESULTS

Themean (x, y) shifts in mm needed to align individually bodies C1 - C5 were respectively (-0.4, 0.5), (+.5, -0.2), (+-0.2, -0.2), (-0.2, +0.4), and (-0.5, +0.7), the monophasic pattern showing acquired curvature along both axes during treatment and demanding a PTV for coverage. A PTV was constructed using a described 6D ellipsoidal based boundary point expansion aligned along the reference frame axis or using a new theory to align against the principal components of the motion. A cyclical one-out method was used to validate the PTV models. Selected confidence intervals yielded complete coverage in >80% weeks in 80% cases. Validation testing disclosed similar complete coverage in 83-86% weekly CBCTs in the test cases with either method. The PCA 6D PTV could yield less normal structure overlap. A one out method was used to test overlap avoidance from PTVs constructed from a population of weekly CBCTs drawn from seven cases with one excluded. PTVs were drawn around target and constrictors on an extraneous case and imaged on a CT slice. Both a rolling 'ball' expansion of the vertices that applies a 3D translational ellipsoid and a PTV constructed using a 6D ellipsoid aligned against the standard reference frame overlapped with all or nearly all the constrictors in all but one trial (1/7). The 6D ellipsoid aligned against the principal motion components spared >70% of a constrictor in all trials (7/7).

CONCLUSION

PTVs remain needed to ensure target coverage in head and neck radiotherapy even with daily CT accuracy because of acquired spinal curvatures resulting in rotational displacements. A described 6D ellipsoid oriented to the reference frame can yield good coverage, but with unneeded constrictor coverage. A PCA analysis yields a PTV with equally good coverage but able to spare 70% of a constrictor.

Machine learning-based prediction of surgical benefit in borderline resectable and locally advanced pancreatic cancer

INTRODUCTION

Surgery represents a primary therapeutic approach for borderline resectable and locally advanced pancreatic cancer (BR/LAPC). However, BR/LAPC lesions exhibit high heterogeneity and not all BR/LAPC patients who undergo surgery can derive beneficial outcomes. The present study aims to employ machine learning (ML) algorithms to identify those who would obtain benefits from the primary tumor surgery.

METHODS

We retrieved clinical data of patients with BR/LAPC from the Surveillance, Epidemiology, and End Results (SEER) database and classified them into surgery and non-surgery groups based on primary tumor surgery status. To eliminate confounding factors, propensity score matching (PSM) was employed. We hypothesized that patients who underwent surgery and had a longer median cancer-specific survival (CSS) than those who did not undergo surgery would certainly benefit from surgical intervention. Clinical and pathological features were utilized to construct six ML models, and model effectiveness was compared through measures such as the area under curve (AUC), calibration plots, and decision curve analysis (DCA). We selected the best-performing algorithm (i.e., XGBoost) to predict postoperative benefits. The SHapley Additive exPlanations (SHAP) approach was used to interpret the XGBoost model. Additionally, data from 53 Chinese patients prospectively collected was used for external validation of the model.

RESULTS

According to the results of the tenfold cross-validation in the training cohort, the XGBoost model yielded the best performance (AUC = 0.823, 95%CI 0.707-0.938). The internal (74.3% accuracy) and external (84.3% accuracy) validation demonstrated the generalizability of the model. The SHAP analysis provided explanations independent of the model, highlighting important factors related to postoperative survival benefits in BR/LAPC, with age, chemotherapy, and radiation therapy being the top three important factors.

CONCLUSION

By integrating of ML algorithms and clinical data, we have established a highly efficient model to facilitate clinical decision-making and assist clinicians in selecting the population that would benefit from surgery.

Clinical concept and relation extraction using prompt-based machine reading comprehension

OBJECTIVE

To develop a natural language processing system that solves both clinical concept extraction and relation extraction in a unified prompt-based machine reading comprehension (MRC) architecture with good generalizability for cross-institution applications.

METHODS

We formulate both clinical concept extraction and relation extraction using a unified prompt-based MRC architecture and explore state-of-the-art transformer models. We compare our MRC models with existing deep learning models for concept extraction and end-to-end relation extraction using 2 benchmark datasets developed by the 2018 National NLP Clinical Challenges (n2c2) challenge (medications and adverse drug events) and the 2022 n2c2 challenge (relations of social determinants of health [SDoH]). We also evaluate the transfer learning ability of the proposed MRC models in a cross-institution setting. We perform error analyses and examine how different prompting strategies affect the performance of MRC models.

RESULTS AND CONCLUSION

The proposed MRC models achieve state-of-the-art performance for clinical concept and relation extraction on the 2 benchmark datasets, outperforming previous non-MRC transformer models. GatorTron-MRC achieves the best strict and lenient F1-scores for concept extraction, outperforming previous deep learning models on the 2 datasets by 1%-3% and 0.7%-1.3%, respectively. For end-to-end relation extraction, GatorTron-MRC and BERT-MIMIC-MRC achieve the best F1-scores, outperforming previous deep learning models by 0.9%-2.4% and 10%-11%, respectively. For cross-institution evaluation, GatorTron-MRC outperforms traditional GatorTron by 6.4% and 16% for the 2 datasets, respectively. The proposed method is better at handling nested/overlapped concepts, extracting relations, and has good portability for cross-institute applications. Our clinical MRC package is publicly available at https://github.com/uf-hobi-informatics-lab/ClinicalTransformerMRC.

Oral mucosal graft ureteroplasty versus ileal ureteric replacement: a meta-analysis

OBJECTIVES

To describe outcomes of oral mucosal graft ureteroplasty (OMGU) and ileal ureter replacement (IUR) and determine the relative merits of both procedures.

METHODS

Databases (including PubMed, Embase and Cochrane) were interrogated for eligible trials that assessed outcomes of OMGU or IUR from 2000 to 30 July 2022. The variables analysed were reconstruction success rates, stricture length, hospital stays, perioperative complications and long-term complications.

RESULTS

A total of 23 single-arm studies were included. The pooled reconstruction success rates for OMGU and IUR were 94.9% (95% confidence interval [CI] 91.0%-97.7%) and 85.8% (95% CI 81.0%-90.0%), respectively. Stricture length of patients in the OMGU and IUR groups were 3.73 (95% CI 3.17-4.28) and 11.55 (95% CI 9.82-13.29) cm, respectively. The maximal stricture length repaired by OMGU was 8 cm. The hospital stays were 5.85 (95% CI 3.88-7.82) and 11.55 (95% CI 6.93-16.17) days in the OMGU and IUR groups, respectively. The incidences of low-grade postoperative complications were 13.6% (95% CI 6.9%-20.3%) and 27.3% (95% CI 19.5%-35.1%), high-grade postoperative complications were 4.6% (95% CI 1.8I-8.5%) and 13.0% (95% CI 9.4%-17.1%), and long-term complications (occurred at > 3months) were 9.0% (95% CI 1.7%-20.0%) and 35.4% (95% CI 25.8%-45.6%) in the OMGU and IUR groups, respectively.

CONCLUSION

An OMGU is an effective, minimally invasive, and safe alternative to IUR for the management of long ureteric strictures. OMGU was the preferred treatment for long ureteric strictures, especially obstructed ureter segments of ≤8 cm.

Contextualized medication information extraction using Transformer-based deep learning architectures

OBJECTIVE

To develop a natural language processing (NLP) system to extract medications and contextual information that help understand drug changes. This project is part of the 2022 n2c2 challenge.

MATERIALS AND METHODS

We developed NLP systems for medication mention extraction, event classification (indicating medication changes discussed or not), and context classification to classify medication changes context into 5 orthogonal dimensions related to drug changes. We explored 6 state-of-the-art pretrained transformer models for the three subtasks, including GatorTron, a large language model pretrained using > 90 billion words of text (including > 80 billion words from > 290 million clinical notes identified at the University of Florida Health). We evaluated our NLP systems using annotated data and evaluation scripts provided by the 2022 n2c2 organizers.

RESULTS

Our GatorTron models achieved the best F1-scores of 0.9828 for medication extraction (ranked 3rd), 0.9379 for event classification (ranked 2nd), and the best micro-average accuracy of 0.9126 for context classification. GatorTron outperformed existing transformer models pretrained using smaller general English text and clinical text corpora, indicating the advantage of large language models.

CONCLUSION

This study demonstrated the advantage of using large transformer models for contextual medication information extraction from clinical narratives.

Identify diabetic retinopathy-related clinical concepts and their attributes using transformer-based natural language processing methods

BACKGROUND

Diabetic retinopathy (DR) is a leading cause of blindness in American adults. If detected, DR can be treated to prevent further damage causing blindness. There is an increasing interest in developing artificial intelligence (AI) technologies to help detect DR using electronic health records. The lesion-related information documented in fundus image reports is a valuable resource that could help diagnoses of DR in clinical decision support systems. However, most studies for AI-based DR diagnoses are mainly based on medical images; there is limited studies to explore the lesion-related information captured in the free text image reports.

METHODS

In this study, we examined two state-of-the-art transformer-based natural language processing (NLP) models, including BERT and RoBERTa, compared them with a recurrent neural network implemented using Long short-term memory (LSTM) to extract DR-related concepts from clinical narratives. We identified four different categories of DR-related clinical concepts including lesions, eye parts, laterality, and severity, developed annotation guidelines, annotated a DR-corpus of 536 image reports, and developed transformer-based NLP models for clinical concept extraction and relation extraction. We also examined the relation extraction under two settings including 'gold-standard' setting-where gold-standard concepts were used-and end-to-end setting.

RESULTS

For concept extraction, the BERT model pretrained with the MIMIC III dataset achieve the best performance (0.9503 and 0.9645 for strict/lenient evaluation). For relation extraction, BERT model pretrained using general English text achieved the best strict/lenient F1-score of 0.9316. The end-to-end system, BERT_general_e2e, achieved the best strict/lenient F1-score of 0.8578 and 0.8881, respectively. Another end-to-end system based on the RoBERTa architecture, RoBERTa_general_e2e, also achieved the same performance as BERT_general_e2e in strict scores.

CONCLUSIONS

This study demonstrated the efficiency of transformer-based NLP models for clinical concept extraction and relation extraction. Our results show that it's necessary to pretrain transformer models using clinical text to optimize the performance for clinical concept extraction. Whereas, for relation extraction, transformers pretrained using general English text perform better.

Prediction of nitrogen excretion from data on dairy cows fed a wide range of diets compiled in an intercontinental database: A meta-analysis

Manure nitrogen (N) from cattle contributes to nitrous oxide and ammonia emissions and nitrate leaching. Measurement of manure N outputs on dairy farms is laborious, expensive, and impractical at large scales; therefore, models are needed to predict N excreted in urine and feces. Building robust prediction models requires extensive data from animals under different management systems worldwide. Thus, the study objectives were (1) to collate an international database of N excretion in feces and urine based on individual lactating dairy cow data from different continents; (2) to determine the suitability of key variables for predicting fecal, urinary, and total manure N excretion; and (3) to develop robust and reliable N excretion prediction models based on individual data from lactating dairy cows consuming various diets. A raw data set was created based on 5,483 individual cow observations, with 5,420 fecal N excretion and 3,621 urine N excretion measurements collected from 162 in vivo experiments conducted by 22 research institutes mostly located in Europe (n = 14) and North America (n = 5). A sequential approach was taken in developing models with increasing complexity by incrementally adding variables that had a significant individual effect on fecal, urinary, or total manure N excretion. Nitrogen excretion was predicted by fitting linear mixed models including experiment as a random effect. Simple models requiring dry matter intake (DMI) or N intake performed better for predicting fecal N excretion than simple models using diet nutrient composition or milk performance parameters. Simple models based on N intake performed better for urinary and total manure N excretion than those based on DMI, but simple models using milk urea N (MUN) and N intake performed even better for urinary N excretion. The full model predicting fecal N excretion had similar performance to simple models based on DMI but included several independent variables (DMI, diet crude protein content, diet neutral detergent fiber content, milk protein), depending on the location, and had root mean square prediction errors as a fraction of the observed mean values of 19.1% for intercontinental, 19.8% for European, and 17.7% for North American data sets. Complex total manure N excretion models based on N intake and MUN led to prediction errors of about 13.0% to 14.0%, which were comparable to models based on N intake alone. Intercepts and slopes of variables in optimal prediction equations developed on intercontinental, European, and North American bases differed from each other, and therefore region-specific models are preferred to predict N excretion. In conclusion, region-specific models that include information on DMI or N intake and MUN are required for good prediction of fecal, urinary, and total manure N excretion. In absence of intake data, region-specific complex equations using easily and routinely measured variables to predict fecal, urinary, or total manure N excretion may be used, but these equations have lower performance than equations based on intake.

Ectopic expression of BrIQD35 promotes drought stress tolerance in Nicotiana benthamiana

The plant IQD gene family is responsive to a variety of stresses. In this study, we studied the structural features and functions of the gene BrIQD35 in Chinese cabbage, a member of the IQD gene family. BrIQD35 was cloned and shown to contain an IQ motif. Transient expression of BrIQD35 indicated that it was localized on the plasma membrane and was significantly upregulated under drought and salt stress in Chinese cabbage. To further identify the function of BrIQD35, it was heterologously overexpressed in Nicotiana benthamiana. Although there was no significant difference between BrIQD35-overexpressed and wild-type (WT) plants under salt stress, WT N. benthamiana showed more wilting than the BrIQD35-overexpressed plants under drought stress. Since the IQ motif has been annotated as a CaM binding site, yeast two-hybrid assays were used to explore the interaction between BrIQD35 and CaM. The results indicated that BrIQD35 interacts weakly with CaMb, but not with CaMa, suggesting that BrIQD35 may function through the Ca2+ -CaMb pathway. The findings reveal a novel gene involved in drought tolerance, which is important for plant breeding and quality improvement for Chinese cabbage.

Susceptibility-weighted imaging cannot distinguish radionecrosis from recurrence in brain metastases after radiotherapy: a comparison with high-grade gliomas

AIM

To explore the efficiency of susceptibility-weighted imaging (SWI) in the differential diagnosis of recurrence from radionecrosis in brain metastases (BM) and in high-grade gliomas (HGG).

MATERIALS AND METHODS

From September 2016 to November 2018, 56 patients with BM and 42 patients with HGG were included in this retrospective study. BM and HGG were assigned to the recurrence and radionecrosis groups according to their histopathology or follow-up results. The proportion of dark signal intensity (proDSI), which was defined as the area of dark signal on SWI or the enhancing area on contrast-enhanced T1-weighted imaging (T1WI), was calculated for each patient. Analysis of variance (ANOVA) with Tukey's honestly significant difference test was used for the repeat multiple comparisons. Receiver operating characteristic curve analysis was performed to validate the diagnostic performance.

RESULTS

For HGG, the proDSI in the recurrence group was significantly lower than that in the radionecrosis group (0.13 ± 0.05 versus 0.43 ± 0.11, p<0.001); however, for BM, no statistical difference was found between groups (0.49 ± 0.09 versus 0.46 ± 0.08, p=0.26). proDSI had the best diagnostic performance (AUC = 0.87, 95% CI: 0.76-0.98; sensitivity = 0.87; specificity = 0.88) for HGG, when a cut-off value of 0.21 was selected.

CONCLUSIONS

Semi-quantitative analysis using SWI is feasible for the differential diagnosis between recurrence and radionecrosis in HGG, but is not feasible in BM. Semi-quantitative assessment based on SWI should interpreted with caution in BM after radiotherapy in clinical practice.

[Analysis of 2 diagnostic criteria of echocardiography for coronary artery aneurysm in Kawasaki disease]

Objective: To analyze the difference between Z score and previous criteria in the diagnosis characteristics of coronary artery aneurysm (CAA) in Kawasaki disease, and to investigate the clinical distribution of Kawasaki disease CAA in the Z score group. Methods: This study retrospectively analyzed the clinical and echocardiographic data of 2 419 children with Kawasaki disease in Shenzhen Children's Hospital from January 2009 to December 2019. The traditional criteria and Z score criteria were used to diagnose CAA, and the differences of diagnostic efficiency between the 2 diagnostic methods were analyzed. The clinical distribution characteristics of CAA in children with Kawasaki disease were analyzed by grouping their sex, clinical classification (complete Kawasaki disease, incomplete Kawasaki disease) the sensitivity to intravenous immunoglobulin (IVIG) (IVIG-sensitive Kawasaki disease,IVIG-unresponsive Kawasaki disease). And the course of the disease (≤6 weeks, >6-8 weeks, >8 weeks to 6 months) etc. The χ² test or Kruskal-Wallis test was used for comparison between the groups, and the Kappa test was used for consistency evaluation. Results: Among the 2 419 children with Kawasaki disease, 1 558 were males and 861 were females. The age of onset was 1.8 (1.0, 3.2) years. The rate of CAA by Z score criteria was higher than that by traditional method (21.9% (529/2 419) vs. 13.9% (336/2 419), χ²=1 074.94, P<0.001). Compared to the traditional method, the Z score criteria found higher rate of CAA in male patients, patients with incomplete Kawasaki disease, and IVIG-unresponsive patients (25.2% (392/1 558) vs. 16.0% (249/1 558), (32.7% (166/507) vs. 19.5% (99/507), 30.5% (95/312) vs. 24.0% (75/312), χ²=694.05, 216.19, 184.37, all P<0.001). The Z score criteria was consistent with the traditional method in diagnosing CAA (κ=0.642,P<0.001). Moreover, in the Z score criteria, the rate of CAA in males (25.2%, 392/1 558) was higher than that in females (15.9%, 137/861), higher in incomplete Kawasaki cases (32.7%, 166/507) than that in complete Kawasaki case (19.0%, 363/1 912), and higher in IVIG-unresponsive cases (30.4%, 95/312) than that in IVIG-sensitive cases (20.6%, 434/2 107), with statistically significant differences (χ²=27.76, 44.38, 15.43, all P<0.001). Coronary Z score of course ≤ 6 weeks was greater than that of course between>6-8 weeks and >8 weeks to 6 months (1.3 (0.7, 2.3) vs. 0.7 (0.3, 1.4), 0.7 (0.3, 1.3), Z=20.65, 13.70, both P<0.001). Conclusions: The rate of CAA in Kawasaki disease by Z score criteria is higher than that by traditional method. In the Z score group, most CAA occur within 6 weeks of the course of the disease, and the rate of CAA in male, incomplete Kawasaki disease, and IVIG-unresponsive is higher.

[Research advances on application of botulinum toxin type A in scar prevention and treatment]

The wound healing time, tension of wound edge, proliferation of fibroblast, and extracellular matrix deposition are the important factors of scar formation, and botulinum toxin type A can regulate the above. Prevention and treatment of scar with botulinum toxin type A is one of the hot topics of clinical research in recent years. This paper briefly reviews researches by scholars at home and abroad on the mechanism, clinical application, complications, and adverse effects of botulinum toxin type A in scar prevention and treatment.

Assessing the Documentation of Social Determinants of Health for Lung Cancer Patients in Clinical Narratives

Social determinants of health (SDoH) are important factors associated with cancer risk and treatment outcomes. There is an increasing interest in exploring SDoH captured in electronic health records (EHRs) to assess cancer risk and outcomes; however, most SDoH are only captured in free-text clinical narratives such as physicians' notes that are not readily accessible. In this study, we applied a natural language processing (NLP) system to identify 15 categories of SDoH from a total of 10,855 lung cancer patients at the University of Florida Health. We aggregated the SDoH concepts into patient-level and assessed how each of the 15 categories of SDoH were documented in cancer patient's notes. To the best of our knowledge, this is one of the first studies to examine the documentation of SDoH in clinical narratives from a real-world lung cancer patient cohort. This study could guide future studies to better utilize SDoH information documented in clinical narratives.

[Evaluation of the perioperative period and long-term outcomes of minimally invasive LTE and minimally invasive CTLE esophagectomy for stage Ⅰ-Ⅲ cervical esophageal carcinoma based on propensity score matching analysis]

Objective: To evaluate the perioperative period and long-term effects of minimally invasive gasless laparoscopic transhiatal esophagectomy (LTE) and minimally invasive combined thoracoscopic and laparoscopic esophagectomy (CTLE) for stageⅠ-Ⅲ cervical esophageal cancer. Methods: The clinical data of 158 consecutive patients with cervical esophageal cancer stageⅠto Ⅲ who underwent minimally invasive CTLE or LTE esophagectomy in the Department of Thoracic Surgery, Beijing Tongren Hospital from January 2008 to December 2019 were retrospectively analyzed. A total of 40 pairs of cases were matched (40 cases of CTLE and 40 cases of LTE surgery) after using the propensity score matching analysis which aimed to balance the influence of confounding factors between groups, including 43 males and 37 females, aged 51 to 81 (62.5±7.0) years old. The perioperative variables and long-term outcomes of the two groups were compared. Results: The operation time ((148.0±31.3) min vs (201.3±48.3) min), intraoperative blood loss ((192.6±77.9) ml vs (387.8±112.4) ml), ICU monitoring time (0 day vs 1 day), and the complication rates of postoperative pneumonia (0 vs 15%) and arrhythmia (2.5% vs 20%) were lower in the LTE group than that of in the CTLE group(all P<0.05). The number of lymph node dissections in the CTLE group was higher than that of in the LTE group (21.2±6.1 vs 12.9±4.3, P<0.001). The 3-and 5-year overall survival (OS) rate and disease-free survival (DFS) rate in the LTE group (OS: 53.53% and 34.27%, DFS: 43.62% and 24.89%, respectively) and the CTLE group (OS: 59.48% and 37.29%, DFS: 49.12% and 28.82%, respectively) had no statistical differences (all P>0.05). Conclusion: The LTE group has advantages in reducing operation time, intraoperative bleeding, ICU monitoring time, postoperative incidence of pneumonia and arrhythmia, and its long-term prognosis is comparable to that of the CTLE group.

TREATMENT OF SECONDARY LOWER LIMB LYMPHEDEMA AFTER GYNECOLOGIC CANCER WITH COMPLEX DECONGESTIVE THERAPY

Secondary lower extremity lymphedema is a common complication of treatment for gynecological cancers. Conservative therapy plays an important role in the treatment of patients with secondary lower extremity lymphedema; in particular, complex decongestive therapy (CDT) has been recognized as an effective nonoperative technique for these patients. But CDT therapy for secondary lower extremity lymphedema remains a problem in China because this technique and its effectiveness have not achieved widespread use and popularity. Our goal was to assess effects of CDT in patients with secondary lower limb lymphedema after treatment for gynecological cancers. The retrospective study consisted of 60 patients who were treated with 20 sessions of CDT. Assessments included objective changes in limb circumference, degree of LE, imaging features, and incidence of erysipelas before and after CDT treatment. We found that CDT can effectively improve lymph stasis and promote backflow, and decrease circumference, interstitial fluid content, and incidence of erysipelas of lymphedematous lower limb. Our results demonstrate that CDT is an effective treatment method for patients with secondary lower limb lymphedema following treatment for gynecologic cancers. This technique should be more widely utilized and popularized in China to improve the quality of life of millions of patients with secondary lower limb lymphedema.

A high precision phase measurement system implemented in FPGA with phase interpolator

High precision timing distribution is crucial to many large scale cosmology and particle physics experiments. Besides the space and energy information, the accurate timing provides an extra dimension for physics event reconstruction. In the timing distribution system, accurate clock phase measurement is an indispensable tool to monitor the phase drift and to achieve accurate phase adjustment. This paper introduces a novel phase measurement method implemented in the Xilinx Field Programmable Gate Array (FPGA). It uses the dedicated phase interpolator in the multi-gigabit transceiver. A design based on this method is implemented within the Kintex Ultrascale series FPGA. The preliminary test result shows that a sub-picosecond level precision is achieved. With this system, the nonlinearity of the phase adjustment in the Xilinx transceiver is measured.

[Association between serum uric acid to creatinine ratio and metabolic syndrome based on community residents in Chashan town, Dongguan city]

Objective: To analyze the association between serum uric acid to creatinine ratio (SUA/Cr) and metabolic syndrome among community residents in Chashan town, Dongguan city. Methods: Participants were from the prospective cohort study of chronic diseases in natural populations in South China conducted in Chashan town, Dongguan city from 2018 to 2019. A total of 11 334 participants with complete data were included by using convenient sampling method. Demographic characteristics, lifestyle and health status were collected through questionnaire and physical examination. The venous blood of the subjects was collected to detect the levels of serum uric acid, creatinine and blood lipid. All participants were divided into four groups (Q1-Q4) according to the quartile of SUA/Cr level. The relationship between SUA/Cr and metabolic syndrome and its components (abdominal obesity, high triglyceride, low level of high density lipoprotein cholesterol, hypertension and abnormal glucose metabolism) were analyzed by using logistic regression model. Results: The mean age of 11 334 participants was (49.52±10.02) years. Male participants accounted for 44.2% (5 015/11 334). The prevalence of metabolic syndrome was 31.2% (3 532/11 334), and the level of SUA/Cr was 5.17±1.53. The prevalence of metabolic syndrome in group Q1-Q4 was 22.3% (631/2 834), 26.5% (752/2 833), 34.9% (988/2 833) and 41.0% (1 161/2 834), respectively. After adjusting for relevant confounding factors, the result of logistic regression model showed that compared with group Q1, the risk of metabolic syndrome in group Q2-Q4 was significantly higher, with OR (95%CI) values about 1.41 (1.23-1.60), 2.19 (1.93-2.49) and 3.01 (2.65-3.42) respectively. The risk of each component of metabolic syndrome in group Q2-Q4 was higher (P_trend<0.001). The SUA/Cr level of participants with normal uric acid level was significantly positively correlated with metabolic syndrome. The risk of metabolic syndrome increased with the increase of SUA/Cr level, but there was the same trend without significant differences in patients with hyperuricemia (P_interaction=0.008). Conclusion: There is a positive correlation between SUA/Cr level and the risk of metabolic syndrome among community residents in Chashan town, Dongguan city.